Samsung SDI's one-line statement in the semi-annual report about "mass production in 27" was read in conjunction with a KAIST research paper — the paper, which the company funded in terms of materials and research expenses, contains remaining challenges.

A KAISTKorea Advanced Institute of Science and Technology
University · South Korea · Website research team measured how the material selected by the company for its solid-state battery, which replaces liquid electrolyte with solid, reacts to humidity inside the factory. We compared the paper with statements the company made in regulatory filings and the newsroom, and with records from the Korean Intellectual Property OfficeKorean Intellectual Property Office
Regulator · Website.

This is not an article evaluating the company, but one checking the stages of evidence. All evidence can be verified through original source links.

Now familiar battery fire news often shows the flammable side to be the liquid electrolyte. Therefore, the solid material replacing the liquid in the solid-state battery is called solid electrolyte. The solid electrolyte selected by Samsung SDISamsung SDIKOSPI 006400
Filings DART · Quote finance.naver.com is a composition made with sulfur, known as sulfide-based.

The company stated in its semi-annual report, which outlines the status of its first-half business, that it plans to mass-produce this battery in 2027. However, a KAIST research team, which received the same type of powder from Samsung SDI, published a paper in Nature CommunicationsNature Communications
Journal · Website before that. The paper examined the surface of the powder, not the company's cell or production line.

What the paper measured is one fact: even inside the dry room where a battery plant builds its cells, the so-called dry room, the surface of this material changed within three days. Measured by the dew point, the temperature at which droplets begin to form, that room sat below minus 60 degrees. Over those three days the ease with which ions pass through, the ionic conductivity, fell by more than a third. This paper is not in the company's filings, and nothing requires it to be. Reading the paper and the company's records side by side does show what homework sits behind the line "mass production in '27", and how to read it is written below.

Read the company records and the paper.

paper

Kim Y-S, Yi J-D, Sung S, Seo D-H. Moisture-induced surface degradation mechanism of argyrodite Li₆PS₅Cl under dry-room conditions. Nature Communications 17, 9270 (2026). Received 2025-09-22, published 2026-07-31. 10.1038/s41467-026-75537-0 · License CC BY-NC-ND 4.0 (this text may be briefly cited but not translated or reproduced) · preprint Research Square 2025-09-12

author affiliations

KAIST Department of Materials Science and Engineering (first author, corresponding author), Korea Automotive Technology InstituteKorea Automotive Technology Institute
Research institute · South Korea · Website, Samsung SDI (Suwon). Samsung SDI-affiliated authors are listed as "assisted in experimental works" in the contribution section.

funding/materials

Funding: "This research was supported by Samsung SDI and the National Research Foundation of KoreaNational Research Foundation of Korea
Funder · South Korea · Website". Methods: The samples were "stoichiometric Li<1>6PS<2>5Cl powders provided by Samsung SDI", and the diffraction and impedance data were "All of the analysis data was given by Samsung SDI". The item that reveals whether the researcher is entangled with the research subject and money is called a conflict of interest declaration, and this paper wrote "no competing interests".

Company records

Semi-annual report (2026.06) II. Content of business · Future business and management plan fair disclosure 2026-07-03 (a disclosure where the company voluntarily reveals undetermined future plans) · Same disclosure 2026-07-02 · Newsroom 2026-07-07 "From robots to electric vehicles... next-generation 'solid-state batteries'

Solid-state batteries appear in company records in two sentences and an investment plan

In the semi-annual report, solid-state batteries appear in only two sentences. In the current status by business division, it states, "The next-generation technology of solid-state batteries aims to lead the industry with mass production by '27." In the research and development section, it says the next-generation technology will be developed through "in-house R&D and strengthened collaboration with overseas research institutes and domestic and international research organizations." The ASB commercialization team under the large-scale business division was newly established in December 2023, according to the business report.

The place where the amount of money is mentioned is a disclosure called fair disclosure, where the company voluntarily discloses its undetermined future plans. Samsung SDI issued two such disclosures on July 2 and 3. The content is that it will spend a total of 25,000,000,000,000 won until 2040 for the Cheonan Madera line, research and development, and the mass production investment of solid-state, LFP, and sodium batteries in Ulsan. Both are attached with the note that "the scale and schedule may change depending on future market conditions and the company's management environment."

The 7 July newsroom article, which contains more technical discussion than regulatory filings, explains the solid electrolyte by categorizing it into sulfide-based, oxide-based, and polymer-based types, then reveals the company's chosen path. The Samsung SDI newsroom also mentions a anode-free technology in which the anode material is entirely omitted, allowing lithium to accumulate spontaneously during charging.

황화물계 전고체 배터리 기술에 집중해 온 삼성SDI는 전고체 배터리의 성능을 끌어올리는데 주력하고 있다. 삼성SDI는 자체 조성한 고체 전해질 소재와 혁신적인 무음극 기술을 전고체 배터리 '솔리드스택(SolidStack)'에 적용했다.삼성SDI 뉴스룸, 2026-07-07

The same article also described the factory's progress, from the pilot line to customer testing.

2023년 3월 국내 배터리 업계 최초로 전고체 파일럿 라인을 구축했다. 같은 해 말부터 시제품 생산에 돌입했으며, 현재는 주요 고객사에 샘플을 공급해 테스트를 진행하고 있다. 이와 함께 제조 공정 기술 고도화와 공급망 구축 등 전고체 배터리 양산을 위한 준비도 병행하고 있다.같은 글

The paper is absent from both the regulatory filing and the newsroom.

I checked if the paper is recorded anywhere in the company's documents. I read the titles of 23 regulatory filings by Samsung SDI uploaded to DARTDART (Data Analysis, Retrieval and Transfer System)
Data source · Website from June to early September, and the full texts of four of them mentioning solid-state technology. I also read nine articles tagged with solid-state technology in the newsroom. The article from July 7 included the full text, but neither this paper nor the research with KAIST appeared anywhere. It is not unusual for a listed company to have no obligation to disclose a paper published jointly with a university, but it means that to find this paper, one must look beyond the filings.

It degrades in three days even with the dry room air.

Sulfide-based solid electrolytes are materials that degrade and emit hydrogen sulfide odor when they come into contact with water, so battery factories produce cells in rooms with extremely low humidity. The question posed in this paper is simple: does this material degrade even in a dry room of the grade used in industrial settings, and if so, in what order does it degrade?

The answer is clear. When the powder provided by Samsung SDI was left in the circulating dry room air, the ionic conductivity dropped by more than one-third within three days. This powder is a representative composition of the sulfide-based material described in the company's newsroom, known as thiogallate in terms of its crystal structure. A few days later, the X-ray diffraction peaks began to broaden. On the surfaces of the samples taken after five and twenty-five days, substances where sulfur had been converted to oxygen increased.

The new finding is not the existence of the problem, but the sequence.

The fact that sulfide electrolytes are vulnerable to moisture is not new. What distinguishes this paper from other studies is that it traces the "why" down to the atomic level. Calculations that directly solve the motion of atoms and electrons from quantum mechanical equations are called first-principles calculations, or DFT in English. The research team tracked water molecules from the moment they land on the surface from dry room air using these calculations and molecular dynamics.

Since calculations alone cannot confirm that this occurred in the actual powder, the research team conducted multiple experiments, gradually etching the surface with argon ions to measure the composition at different depths. This analysis, which reads the composition from electrons emitted from the surface, is called X-ray photoelectron spectroscopy, abbreviated as XPS. The order in which the calculations and experiments aligned is as follows.

1. Water molecules adhere to surfaces rich in sulfur.
2. The indium–sulfur bonds become weaker, and it becomes thermodynamically favorable for oxygen to enter in place of sulfur.
3. Oxygen-containing PS₄ tetrahedra rotate, pushing oxygen deeper into the surface.
4. The surface transforms into a layer similar to Li₆PO₅Cl, rich in oxygen.
5. This layer breaks down in a way that reduces volume, forming LiCl, Li₃PO₄, Li₂SO₄, LiOH, and Li₂CO₃.

The final stage is critical. Usually, the oxide layer formed on the surface acts as a protective shell for the interior. This self-generated shell that prevents further degradation is called passivation. However, in this case, the volume of the product is smaller than the original, causing holes in the shell and allowing water to continuously penetrate inward. The authors wrote that this layer "fails to passivate", meaning it does not form passivation, and thus the degradation does not stop.

The direction of the solution also comes from the same place. The authors proposed using a composition that prevents tetrahedral rotation in the third step, but they suggested doping, which involves adding only a small amount of other elements to the original composition, in very small quantities so as not to block the path for ions. They added that one should choose elements that produce large-volume byproducts when decomposed, so that the shell can cover without holes. The authors conclude in the final paragraph that following this direction increases the likelihood of compatibility with "industrial dry-room environments". This is a proposal, not something that has been created and verified.

The places where the paper and the company's statements diverge are threefold.

When we compare the company records and the paper, what we always check is three things, not to say which side is right or wrong, but to measure the distance between what the paper shows and what the company says.

Looking at who wrote the paper, it is awkward to call it either a company paper or an academic paper. The planning, experiments, calculations, and writing were done by the KAIST researcher, while Samsung SDI provided the materials and diffraction and impedance data, along with research funding. One affiliated researcher assisted with the experiments. It is more accurate to read this as a study where the university answered a question posed by the company using the company's materials, and it is noted that there is no declaration of conflict of interest.

The following is a matter of checking whether the paper's subject and the company's product are the same. The paper's samples are powder and pellets, not cells. While the newsroom describes SolidStack as a cell with a no-electrode structure that uses "self-developed solid electrolyte," the material in the paper is "stoichiometric Li<2>6PS<3>5Cl", a textbook composition. Neither the paper nor the company states whether this composition is the same as that used in the company's cells. The doping proposed in the paper may or may not already be included in the company's composition, but neither record provides an answer.

What remains is the conclusion sentence, where the company's statement is "aiming for mass production in '27." The conclusion of the paper states that this material has "practical operating environments" within "critical vulnerability" and the design principle to reduce it. Since the paper does not mention the mass production timeline at all, the only overlapping element is the word "dry room." The company says it is preparing for mass production in that room, while the paper measured how the material changes within that room.

Limitations stated by the authors, limitations visible upon reading

• This is the story of powder and pellets. Inside the cell, the electrolyte is pressed against the cathode and anode, and the surface area exposed to air differs. The figure of 36% over three days cannot be directly translated into the cell's lifespan.
• The diffraction and impedance data were measured and provided by the company. The paper directly measured only the XPS depth analysis. The paper briefly mentions this in the Methods section. It is not something to hide, but the original data behind the 36% figure is outside the university.
• The calculations were performed based on an ideal surface model. The actual powder defects, particle size, and residual impurities are not included in the model. The authors also write that they read the sequence by combining calculations and experiments.
• The solution is a proposal. There is no data in this paper about doping that prevents rotation or elements that create large-volume products, nor about creating and placing them in a dry room.
• This issue is not unique to Samsung SDI. The prior research cited in the introduction comes from university and corporate groups in various countries. Any company choosing a sulfide-based system would use the same room. The paper is a record of one company explaining the principle using its own material.

Method of verification

The full paper can be freely accessed by anyone via the DOI above. The cited sentences are in the last two paragraphs of the Abstract and Results and discussion, in "Material characterization" of Methods, and in Funding and Contributions of Author information. The company's statements are from the second chapter of the DART semi-annual report, "Content of Business," specifically the seventh item, Other Reference Matters, the sixth item, Research and Development Activities, and the main text of the newsroom link. The failure to find the paper in the company's filings is within the scope of our review of the DART list and the newsroom.

Arrange the company records and the paper in chronological order

As in the first post, it is not possible here to collect and count in bulk similar types of disclosures. Korean companies do not disclose papers, and the unique identifiers (DOIs) assigned to each paper are rarely attached to the DART filings, so there is no sample linking papers and stock prices. Instead, what can be done is to arrange in chronological order the records left by the company regarding solid-state technology and the records left by the paper and compare them.

The company filed a patent for the same issue first

The paper was not disclosed in the filing, but the company did not leave this topic unrecorded. There are patents. We searched the Korean Intellectual Property Office database for aziridinate, the material of this paper, and moisture, the issue addressed in this paper. A published patent application is a document that discloses the application content to the public, regardless of examination results, as shown in the table below.

Samsung SDI, which led the list of applicants in the azidoate sample, also submitted this item in the water content category.

수분 안정성이 우수한 고체전해질 및 이의 제조방법삼성에스디아이, 출원 2022-10-13

The paper was submitted to Nature Communications in September 2025. The company filed a patent on the same issue in 2022, three years earlier. This is not to indicate which came first, the paper or the patent, but to show that the company's awareness of the issue is recorded in public filings. It is not disclosed in regulatory filings or the newsroom, but it is in the patent.

Rejected applications are also recorded

Among the 200 samples, 13 cases ended in rejection, and the list includes not only automotive companies and government-funded research institutes but also industry-academia cooperation units, which handle research contracts and patents for universities.

Hyundai and Kia filed the same title in 2017 and 2023 and were rejected both times, but we cannot read the rejection reasons and thus cannot say which side is correct. However, the record shows that in this field, even research institutes of automakers, government-funded research institutions, and university-industry collaboration teams have been rejected. The paper's statement that "existing strategies are insufficient" appears similarly in the examination records.

There are clear limitations. Due to the free search limit, only 100 cases each were selected from a total of 1,372 and 1,768 cases. The applicant rankings and status ratios are based on samples, not the full dataset, and only titles and bibliographic information were reviewed, not the claims — it was not confirmed whether the composition in the paper matches the composition in the patent. The search terms only capture patents where the terms appear in the title or abstract, so patents using different wording are excluded.

Why not talk about stock prices?

July 31, the day the paper was published, was the day after the company disclosed its second-quarter earnings. On the company's IR page, the previous day's closing price of 358,500 won was the lowest during the period on the six-month chart. The closing price on September 7 was 541,000 won, with the company disclosing the end of the GM joint venture and the semi-annual report in between. Since this paper was neither disclosed nor reported, there is no basis to claim the market reacted to the paper, and we do not write it that way.

The 143 U.S. cases seen in the first case were instances where the company voluntarily included the paper in an 8-K filing. This case is the opposite, where the company did not include the paper but found and read it externally, so the extended layer became a comparison of records rather than a market response.

Solid-state batteries are mentioned only in text within the semi-annual report.

The "R&D achievements" table is a section where the company selects and lists major projects from the past three years. Therefore, the absence of solid-state in this table does not imply anything beyond its absence in the table. Solid-state appears in two other sections of the same report (business segment status, R&D activity overview) in sentence form.

In national R&D, this issue is moving toward air-stable cells and mass production

The Funding section at the end of the paper lists four project numbers from the National Research Foundation of Korea along with Samsung SDI. We checked each number one by one at the National Science & Technology Information ServiceNational Science & Technology Information Service
Data source · Website, commonly abbreviated as NTIS. Since NTIS registers one project each year for multi-year projects, it separately lists the number of projects and annual entries.

The four tasks focus on oxygen stability in cathodes, material prediction using machine learning, sodium batteries, and oxygen redox cathodes. No task addresses the issue of moisture in sulfide electrolytes, meaning Samsung SDI is the only one to bet money on the questions raised in this paper. Government tasks should be understood as funding to operate research laboratories. For Professor Dong-Woo Kim of KAIST, the corresponding author, there is also a KAIST-MITMassachusetts Institute of Technology
University · United States · Website joint leading research center listed on NTIS, which receives 2,300,000,000 won annually for individual basic research tasks and data-driven solid electrolyte material design tasks.

Samsung SDI rarely appears on the list of national projects.

Then, how is a company that provides materials and research funding captured in national R&D. Among the 4,035 annual reports retrieved by searching for solid electrolytes, only two are led by Samsung SDI, both of which are solar cell projects from 2008 and 2009 and unrelated to solid-state batteries. There are no national projects on solid-state or sulfide batteries led by Samsung SDI.

However, this does not mean the company is unrelated to national projects. Large corporations rarely appear as lead organizations in national R&D, and participation is not visible in these data because it is not a publicly listed item on the NTIS.

The national funding allocated to sulfide electrolytes has exceeded 100,000,000,000 won cumulatively.

By narrowing the national tasks to sulfide-based solid electrolytes, the scale of government spending on this material becomes clear. New tasks increased from three in 2019 to 29 in 2026, and the total government funding accumulated since 2012 exceeds 100,000,000,000 won.

Among these, I selected only the projects that mention in their titles the issue addressed in this paper: resistance to air and moisture. The oldest is a 2017 KISTKorea Institute of Science and Technology
Research institute · South Korea · Website project, and more recent ones increasingly shift from materials to cells and mass production.

Government investment research funds are listed for the respective year. Projects spanning two years are divided by year. The items listed in the table include only those permitted for external disclosure (project name, implementing organization, research director, project name, research period, government investment research funds).

The term "atmosphere-stable sulfide" already appeared in a 2017 KIST project title. Therefore, what this paper newly revealed is not the existence of the problem, but the atomic-level sequence of degradation and why the protective layer fails. Government funding shifted during that period from materials to atmosphere-stable cells and wet mass production, but whether those projects overlap with the doping direction proposed by the paper cannot be determined from the project titles alone.

So where should one check?

This article does not draw a conclusion but instead lists three points to verify when reading a line such as ``mass production in '27.''

1. Is the company composition the same as the textbook composition? What the paper described is the textbook composition. The company only disclosed that it uses its own composition. Whether the two are the same is not mentioned in either record. One can compare them when the company discloses its composition or doping.
2. The patent appears 1.5 years later. Applications filed after 2025 have mostly not been published yet. By counting the number of applications with "moisture" or "atmospheric stability" in their titles in the 2027 publication list, one can later infer where companies are currently focusing their efforts.
3. Results of the wet mass production project. The area where government funding shifted is the wet mass production of 1 ton per month. When the three companies' projects end, the results will be recorded in the NTIS achievement items and each company's filings.

What this paper did not answer is whether the one-third measured in the powder also remains the same size inside the cell. The answer can only come from the company's cell measurement data. The next verification points are whether the company discloses its composition and cell data, and the 2027 publication list when applications filed after 2025 begin to be published. A single paper is just a single paper, but a paper that a company has funded with materials and money is one of the few publicly available records that indicate what the company is currently solving.